Effect of rainfall on the grain yield of spring wheat, with an application to the analysis of adaptation

Abstract

The grain yields for nine cultivars of spring wheat grown at 11 sites in 5 years in the Central West of New South Wales have been previously analysed. The results suggested that the variability in yield was mainly attributable to a single environmental factor. Four components of annual rainfall are now defined, and it is found that each correlates positively with yield over the 37 trials carried out. The highest correlation is with spring rainfall, defined as the rainfall from 3 weeks before anthesis to 2 weeks after anthesis. Year, site and year x site effects are calculated for yield and for spring rainfall by means of a least-squares procedure. The correlations between rainfall and the mean yield of the nine cultivars are 0.983, 0.947 and 0.782 for year, site and year x site effects respectively, which indicates that spring rainfall or a weather factor closely correlated with spring rainfall is a major determinant of grain yield. An adaptation analysis with spring rainfall as the independent variable is compared with the more conventional analysis involving site mean yield as the independent variable. The former analysis is shown to have the advantage that the relative stabilities of the cultivars can be interpreted in terms of their response to rainfall.